YJIT Benchmarks

Details for Benchmarks at 2024-07-26 19:08:32 UTC

YJIT metrics from the yjit-bench suite using Ruby e798f45cc2

Using the geomean of the headline benchmarks for x86 YJIT 3.4.0dev is
  • 79.8% faster than CRuby 3.4.0dev
  • 3.5% faster than YJIT 3.3.3
On railsbench it is
  • 79.7% faster than CRuby 3.4.0dev
  • the same speed as YJIT 3.3.3

Performance on Headline Benchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on Headline Benchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Performance on Other Benchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on Other Benchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Performance on MicroBenchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on MicroBenchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Want Raw Graphs and CSV?

Benchmarks Speed Details

Benchmark Memory Usage Details

Number of Iterations and Warmups Tested

bench CRuby 3.3.3 warmups CRuby 3.3.3 iters CRuby 3.4.0dev warmups CRuby 3.4.0dev iters YJIT 3.3.3 warmups YJIT 3.3.3 iters YJIT 3.4.0dev warmups YJIT 3.4.0dev iters
activerecord 10 51 10 47 10 116 10 123
chunky-png 10 18 10 14 10 36 10 45
erubi-rails 10 10 10 10 10 20 10 19
hexapdf 10 10 10 10 10 10 10 10
liquid-c 10 347 10 332 10 494 10 491
liquid-compile 10 335 10 329 10 476 10 485
liquid-render 10 136 10 130 10 351 10 367
lobsters 10 14 10 13 10 24 10 23
mail 10 160 10 148 10 233 10 223
psych-load 10 10 10 10 10 10 10 10
railsbench 10 10 10 10 10 10 10 10
rubocop 10 135 10 127 10 227 10 237
ruby-lsp 10 129 10 122 10 207 10 215
sequel 10 318 10 306 10 427 10 423
binarytrees 10 51 10 49 10 118 10 119
blurhash 10 61 10 56 10 119 10 141
erubi 10 83 10 83 10 103 10 113
etanni 10 58 10 58 10 68 10 71
fannkuchredux 10 10 10 10 10 17 10 16
fluentd 10 10 10 10 10 10 10 10
graphql 10 10 10 10 10 10 10 10
graphql-native 10 34 10 22 10 40 10 24
lee 10 11 10 10 10 20 10 20
matmul 10 10 10 10 10 15 10 16
nbody 10 213 10 200 10 410 10 462
nqueens 10 113 10 104 10 114 10 465
optcarrot 10 10 10 10 10 10 10 10
protoboeuf 10 183 10 169 10 603 10 690
protoboeuf-encode 10 123 10 126 10 268 10 345
rack 10 455 10 474 10 817 10 893
ruby-json 10 10 10 10 10 10 10 10
rubykon 10 10 10 10 10 10 10 10
sudoku 10 10 10 10 10 27 10 31
tinygql 10 25 10 21 10 52 10 52
30k_ifelse 10 10 10 10 10 70 10 69
30k_methods 10 10 10 10 10 26 10 26
cfunc_itself 10 255 10 226 10 1099 10 1212
fib 10 107 10 98 10 616 10 618
getivar 10 247 10 193 10 1279 10 1695
keyword_args 10 86 10 81 10 720 10 946
object-new 10 225 10 193 10 291 10 250
respond_to 10 92 10 82 10 1737 10 2090
ruby-xor 10 31 10 25 10 100 10 144
setivar 10 370 10 411 10 2334 10 4873
setivar_object 10 242 10 250 10 646 10 702
setivar_young 10 238 10 261 10 644 10 698
str_concat 10 337 10 298 10 605 10 551
throw 10 949 10 901 10 1244 10 1168

Different Ruby configurations want different amounts of warmup. With no JIT, CRuby needs hardly any. YJIT and MJIT 3.0 both warm up quite quickly, while MJIT in 3.1 often slows down for a time as it compiles, after an unpredictable delay.

Benchmark YJIT Stats

Hover your cursor over the benchmark names for descriptions of each benchmark.

bench Exit Report Inline Outlined Comp iSeqs Comp Blocks Inval Inval Ratio Bind Alloc Bind Set Const Bumps
activerecord (click) 1416074 1263939 214 2225 0 0% 0 0
chunky-png (click) 331819 296083 84 1093 1 0% 0 0
erubi-rails (click) 1715348 1394757 288 3126 17 0% 0 0
hexapdf (click) 1301353 1108477 506 11846 41 0% 0 0
liquid-c (click) 516647 479100 119 1768 5 0% 0 0
liquid-compile (click) 439972 391232 151 2089 2 0% 0 0
liquid-render (click) 613188 559911 136 2333 8 0% 0 0
lobsters (click) 7632218 6416010 3077 50178 86 0% 0 0
mail (click) 777658 720649 347 5643 14 0% 0 0
psych-load (click) 288205 264414 61 594 3 0% 0 0
railsbench (click) 2982834 2481533 1650 16294 71 0% 0 0
rubocop (click) 5517771 4759996 2891 51128 110 0% 4 0
ruby-lsp (click) 840488 711220 310 4687 24 0% 0 0
sequel (click) 480695 390477 16 117 0 0% 0 0
binarytrees (click) 8765 7825 10 70 0 0% 0 0
blurhash (click) 50455 44132 31 457 0 0% 0 0
erubi (click) 269312 214592 10 120 0 0% 0 0
etanni (click) 34400 25222 12 104 0 0% 0 0
fannkuchredux (click) 24067 31522 5 256 0 0% 0 0
fluentd (click) 371098 317678 4 28 0 0% 0 0
graphql (click) 438868 381242 66 650 0 0% 0 0
graphql-native (click) 410103 348604 41 273 0 0% 0 0
lee (click) 293048 215306 34 557 0 0% 0 0
matmul (click) 8318 9199 8 86 0 0% 0 0
nbody (click) 15375 19248 11 195 0 0% 0 0
nqueens (click) 24183 24054 10 283 0 0% 0 0
optcarrot (click) 309371 266843 188 4402 34 0% 0 0
protoboeuf (click) 140353 122267 17 1368 0 0% 0 0
protoboeuf-encode (click) 221919 196362 20 1179 0 0% 0 0
rack (click) 271523 225845 37 413 0 0% 0 0
ruby-json (click) 30520 28961 8 167 0 0% 0 0
rubykon (click) 132662 103297 138 1522 3 0% 0 0
sudoku (click) 45653 24997 8 520 0 0% 0 0
tinygql (click) 304362 273316 58 735 5 0% 0 0
30k_ifelse (click) 5134674 4729436 9259 50778 0 0% 0 0
30k_methods (click) 2016920 1591842 5778 19339 0 0% 0 0
cfunc_itself (click) 7533 6248 9 70 0 0% 0 0
fib (click) 4970 5023 8 49 0 0% 0 0
getivar (click) 5936 6875 8 74 0 0% 0 0
keyword_args (click) 8408 6980 10 72 0 0% 0 0
object-new (click) 4400 4255 7 45 0 0% 0 0
respond_to (click) 8276 8394 9 85 0 0% 0 0
ruby-xor (click) 5580 6003 5 62 0 0% 0 0
setivar (click) 5087 5551 8 54 0 0% 0 0
setivar_object (click) 5470 5459 8 54 0 0% 0 0
setivar_young (click) 6180 6181 9 62 0 0% 0 0
str_concat (click) 7776 8178 11 85 0 0% 0 0
throw (click) 7899 6453 10 69 0 0% 0 0

YJIT stats correspond to the YJIT stats exit report.

Note: currently, all stats are collected on x86_64, not ARM.

Raw JSON data files

All graphs and table data in this page comes from processing these data files, which come from benchmark runs.